Conditional subjective probability density approach to dependency consideration in passive system reliability

Passive safety systems (PSSs) are increasingly being adopted in the design and operations of nuclear installations due to their simplicity, economics, inherent safety and reduced need for human intervention. Several critical parameters (CPs) which influence the performance and consequently reliability of thermal-hydraulic (t-h) nuclear passive systems are continuously been discovered. It is conventionally assumed in most cases that the CPs of reliability are independent but in reality they are dependent and their dependency is stochastic. The issue of dependency is always complicated even for sub-systems of similar characteristics and this necessitates multivariate dependency analysis. Thus, there is need to consider the dependency effects of these influencing parameters of reliability of the t-h nuclear systems prior and after their deployment to guarantee safety and improve performance respectively. Dependency analysis approaches are now being incorporated into reliability methodologies in order to obtain more realistic results. In this paper, a generalized procedure was given for evaluating the reliability of t-h PSSs based on dependency consideration of identified key CPs. One of the possible methods for dependency analysis of CPs - conditional subjective probability density function approach - was presented. A passively water cooled steam generator was used to demonstrate the dependency nature of the identified key CPs by applying the presented method. In addition, the influence of the dependent CPs on reliability was also demonstrated. The results obtained justified the need to consider the dependency effects of the CPs. In addition, to obtain more realistic t-h reliability results, it is better to identify and select as many key CPs as possible and subject them to a suitable multivariate analysis approach such as the conditional subjective probability density function approach.